r - Q: initialization of Genetic Algorithm -


i have function big search space, wanted use genetic algorithms close optimum, , use other method such bfgs find optimal point. i'm using r that.

the problem function has 12 parameters , invalid @ points. when give maximum , minimum parameter values ga , tries generate population, whole population formed nans (and therefore, algorithm not able continue).

since don't know gas, i'm struggling find solution problem. thing thought choosing big population size (something 10e5), find valid values start. that's not solution though, being able initialize algorithm normal-sized population better.

do have suggestions? gas i'm missing?

thank you

if of parameter space invalid, ga may not best approach. risk cross algorithms, may invalid parameter combination.

another way of thinking ga search algorithm. works best when tries ends little bit better or worse started from; if tries invalid there isn't guiding search.

my suggestion try parametrize problem differently, more points valid (though perhaps low-fitness). didn't describe function it's hard more specific, it's possible accept "invalid" input , quantify how far "valid" is. that's can run search on.


Comments

Popular posts from this blog

c++ - Difference between pre and post decrement in recursive function argument -

php - Nothing but 'run(); ' when browsing to my local project, how do I fix this? -

php - How can I echo out this array? -